the aging population increases, so does the incidence of stroke and consequent communication deficits in such individuals. Although many treatment paradigms exist for aphasia rehabilitation with behaviorally successful outcomes, it is still unclear what specific positive neurophysiological changes are occurring as the result of successful treatment. The proposed project will examine the neuroplastic response in persons with aphasia to a theoretically-based generative naming treatment that has been shown to produce positive behavioral results. This study will focus on the processing of abstract and concrete words because they have been shown to be differentially processed both behaviorally (see Paivio, 1991 for a review) and neurally (e.g., Binder et al., 2009) in neurologically healthy adults. This behavioral pattern is exaggerated in persons with aphasia (e.g., Newton & Barry, 1997) and can be manipulated in treatment to produce gains not only in treated (abstract) items, but also generalization to untrained (concrete) items (Kiran, Sandberg, & Abbott, 2009). This treatment will therefore be used in concert with pre- and post- treatment fMRI scans to link changes in activation patterns and effective connectivity within language networks to a positive behavioral response to treatment. Six participants with aphasia will be scanned using fMRI before and after treatment. The word generation treatment that will be used has previously shown that training abstract words in a particular context results in improvement for those words as well as concrete words in the same context, but training concrete words only improves the trained items (Kiran, Sandberg, & Abbott, 2009). Therefore, only abstract words will be trained in the proposed study to examine changes in neural activation patterns for both abstract and concrete words as a function of improved lexical retrieval subsequent to treatment. Traditional fMRI analysis will be used to examine changes in activation patterns for abstract and concrete word processing from pre- to post- treatment. Structural equation modeling will be used to examine changes in connectivity between regions in abstract and concrete word processing networks from pre- to post-treatment. The innovation of the proposed work lies in combination of a) the use of both traditional fMRI and effective connectivity analyses to measure treatment gains in different language networks, b) the attention to specific kinds of stimuli that can engage the brain differently, and c) the use of a theory-based generative naming treatment that has been shown to be successful for treating word retrieval deficits in chronic aphasia. The successful completion of this project will increase our understanding of how treatment influences neural plasticity and will guide how new treatments for word-finding deficits in aphasia will be developed.